AI-based liquid fertilizer drip irrigation control method and system for agricultural planting

Abstract

This invention relates to the field of drip irrigation regulation technology, specifically to an AI-based method and system for regulating liquid fertilizer drip irrigation in agricultural planting. The method includes the following steps: collecting data on pipeline flow rate, drip main flow velocity, and changes in soil moisture and ion concentration in the crop root zone from the liquid fertilizer storage tank; analyzing the delivery differences at different irrigation nodes; controlling the action of zoned solenoid valves according to priority and recording the execution response to form a steady-state regulation range parameter. This invention achieves data fusion of root zone response and pipeline status through multi-channel signal standardization processing of feedback feature vector sets. Node weight distribution information reflects differences in delivery performance, execution time sequence offset signals quantify the consistency of action response, and the steady-state regulation range parameter supports system adaptive optimization, forming a continuous adjustment and closed-loop response mechanism. This allows for flexible responses to changes in crop root zone absorption and fluctuations in pipeline delivery status, improving the coordination and operational reliability of the water and fertilizer regulation system.

Description

Technical Field

[0001] This invention relates to the field of drip irrigation control technology, and in particular to an AI-based method and system for controlling drip irrigation of liquid fertilizers for agricultural planting. Background Technology

[0002] Drip irrigation control involves technical methods for precise control of water and fertilizer supply during agricultural irrigation. It mainly includes key aspects such as the design and layout of the drip irrigation system, the setting and adjustment of the water-fertilizer mixing ratio, the collection and modeling of crop water and fertilizer requirements, the scheduling of irrigation times, and the implementation of control methods. Among these, the traditional drip irrigation control method for liquid fertilizers in agricultural planting refers to the method of periodically delivering liquid fertilizer and water to the root zone of crops by setting irrigation programs and fertilization plans based on crop growth cycles and soil moisture conditions, using timed controllers or logical rules based on human experience.

[0003] Current drip irrigation control methods rely on fixed time programs and human experience, and depend on crop growth cycles and soil information. They lack dynamic feedback collection and multi-parameter linkage judgment of data from multiple points within the drip irrigation network. The control process is slow to respond to differences in the delivery status between irrigation nodes, and it is difficult to detect delivery anomalies or untimely responses during operation. The correlation between pipeline operation status and changes in crop root zone absorption feedback is limited, making it difficult for the control link to adapt to changes in actual demand in a timely manner, affecting the coordination of water and fertilizer supply and irrigation stability. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and propose an AI-based method and system for regulating liquid fertilizer drip irrigation in agricultural planting.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an AI-based drip irrigation control method for liquid fertilizer in agricultural planting, comprising the following steps: S1: Based on the operating data of liquid fertilizer storage tank and fertilizer mixing pump, analyze the changes in flow signal in the pipeline, detect the changes in the flow rate of the drip irrigation main pipe, monitor the soil moisture and ion concentration response in the crop root zone through sensors, identify the periodic signal amplitude characteristics, and obtain the feedback feature vector set; S2: Based on the feedback feature vector set, compare the performance of each parameter among different irrigation nodes, analyze the delivery status of the drip irrigation main node, determine the node association based on the matching degree between the root zone feedback signal and the pipeline status, calculate the delivery difference, and obtain the node weight distribution information. S3: Based on the node weight distribution information, the root zone feedback response of the drip irrigation zone is judged by the near-end strategy optimization algorithm. Combined with the liquid state of the mixed fertilizer pump, the feedback trend of the drip irrigation path is analyzed by comparison. The node scheduling order is adjusted according to the trend to obtain the control priority sequence. S4: Based on the control priority sequence, optimize the action sequence of the partition solenoid valves, issue valve action commands in order, record the valve action response fed back by the controller, analyze the time difference between the command sequence and the response process, and obtain the execution timing offset signal; S5: Based on the execution timing offset signal, filter continuous periodic parameter changes, analyze the stability state of the valve response process, record the control structure state when the response is consistently consistent, and supplement the scheduling record when response fluctuations are detected, so as to obtain the control steady-state interval parameters.

[0006] The present invention is improved in that the feedback feature vector set includes signal amplitude features, signal response balance, and data consistency index within the period; the node weight distribution information includes node transmission efficiency coefficient, correlation strength parameter, and transmission difference distribution; the control priority sequence includes path priority number, feedback trend classification, and node sorting identifier; the execution timing offset signal includes valve action response time difference, trigger sequence matching degree, and execution offset parameter; and the control steady-state interval parameter includes periodic stability discrimination parameter, interval switching identifier, and operating status number.

[0007] The present invention is improved in that the step of obtaining the feedback feature vector set is specifically as follows: S111: Based on the operating data of liquid fertilizer storage tank and mixed fertilizer pump, analyze the changes in the output signal of the flow sensor as the fertilizer solution flows in the pipeline, compare the data sequence synchronously collected by the pressure sensor, calculate the continuous distribution of the flow velocity at each monitoring point over time, screen the time synchronization segments of soil moisture response and ion concentration response in the crop root zone, determine the data mapping of each channel signal at the same time, and obtain multi-channel sensor response data. S112: Based on the multi-channel sensing response data, analyze the fluctuation range of each channel signal within the complete cycle, calculate the amplitude change of each sampling point, identify continuous data segments within the cycle boundary, compare the dynamic responses between different signals, and obtain the channel fluctuation amplitude set. S113: Based on the set of channel fluctuation amplitudes, analyze the overall distribution of amplitude changes in all channels, calculate the feature vectors under a unified scale, determine the data balance in the periodic response, optimize the expression format of cross-channel changes, and obtain the feedback feature vector set.

[0008] The present invention is improved in that the step of obtaining the node weight distribution information is specifically as follows: S211: Based on the feedback feature vector set, analyze the corresponding parameters of each irrigation node, compare the amplitude feature differences between nodes in the same irrigation cycle, determine the signal response changes of each node in the same time period, calculate the spatial distribution range of node parameters, and obtain the irrigation node parameter difference sequence. S212: Based on the difference sequence of irrigation node parameters, analyze the fluid velocity changes of each node, compare the fluid continuous transmission state between nodes, determine the temporal response consistency between the root zone feedback signal and the main fluid state, identify nodes with correlation within the time period, adjust the synchronization structure of the response data, and obtain node synchronization response correlation data. S213: Based on the node synchronous response association data, calculate the offset parameters of the transport state between each node, analyze the coupling strength between the node feedback signal and the fluid transport characteristics, determine the control capability of the master control node, optimize the mapping structure between node number and transport parameters, and obtain node weight distribution information.

[0009] The present invention is improved in that the step of obtaining the control priority sequence is specifically as follows: S311: Based on the node weight distribution information, the operation data of the drip irrigation nodes are analyzed through the near-end strategy optimization algorithm. The humidity response and nutrient concentration changes in the root zone of the irrigation area are combined and processed to determine the dynamic synchronicity of crop root zone water changes and ion concentration fluctuations between drip irrigation paths. The fluid transport status of the irrigation path is synchronously detected to obtain the path feedback trend set. S312: Based on the path feedback trend set, compare it with the fluid transport characteristics during the operation of the fertilizer pump, calculate the normalized result of the fluctuation amplitude of each path feedback and the pipeline pressure response, obtain the feedback trend intensity factor, determine the strength relationship of the feedback trends between each path, and obtain the node control sorting information. S313: Based on the node control and sorting information, prioritize all drip irrigation paths, mark the feedback trend level and associated node identification number, optimize the path arrangement order, and obtain the control priority sequence.

[0010] The present invention is improved in that the step of obtaining the execution timing offset signal is specifically as follows: S411: Based on the control priority sequence, control commands are sent to the solenoid valve in sequence to optimize the action signal issuance process. The issuance time of each command is uniformly recorded using time synchronization, the scheduling order of each command is determined, and a valve command scheduling time sequence set is obtained. S412: Based on the valve command scheduling timing set, synchronously collect the action response time of each solenoid valve fed back by the controller, calculate the time interval between the action response and the command issuance, obtain the response time offset of each solenoid valve, and obtain the response offset vector set. S413: Based on the response offset vector set, filter all signals with sequential offsets, determine the temporal sequence relationship of each group of signals, and archive the offset time and sequence characteristics to obtain the execution timing offset signal.

[0011] The present invention is improved in that the step of obtaining the control steady-state range parameter is specifically as follows: S511: Based on the execution timing offset signal, analyze the response sequence of each valve in the continuous drip irrigation cycle, compare the distribution of the start and end nodes of the valve response in different cycles, determine whether the unfolding order of the response process in each cycle is stable, identify response segments with similar unfolding trajectories, and obtain cycle response stability data. S512: Based on the periodic response stability data, analyze the drip irrigation path response, compare the response order of each path in a continuous period, determine the consistency of the path response order in different periods, identify the path segments whose response order remains continuous, and perform hierarchical identification on the segments where the response order changes, to obtain the drip irrigation path mapping result. S513: Based on the drip irrigation path mapping results, analyze the scheduling and control characteristics within each segment, determine the continuity of the response structure within each path segment, identify scheduling parameters that exhibit consistency in continuous segments, optimize the numbering and scheduling relationship of each segment, and obtain the steady-state interval parameters for regulation.

[0012] The present invention is improved in that the change in flow signal refers to the fluctuation of the electrical signal output by the flow sensor when the liquid fertilizer flows in the pipeline as the flow rate and flow rate of the fertilizer solution change. The irrigation node refers to the location node in the drip irrigation system that can be independently controlled and data collected, including the main pipe, branch pipe and zone valve.

[0013] An AI-based drip irrigation control system for liquid fertilizer in agricultural planting, the system comprising: The feedback acquisition module is based on the liquid fertilizer storage tank, analyzes the changes in flow signals in the pipeline, detects the changes in the flow rate of the drip irrigation main pipe, monitors the response of soil moisture and ion concentration in the crop root zone, identifies the periodic signal amplitude characteristics, and obtains the feedback feature vector set. Based on the feedback feature vector set, the weight analysis module compares the performance of each parameter among different irrigation nodes, analyzes the delivery status of the drip irrigation main node, determines the node association based on the matching degree between the root zone feedback signal and the pipeline status, calculates the delivery difference, and obtains the node weight distribution information. The priority determination module judges the root zone feedback response of the drip irrigation zone based on the node weight distribution information, and analyzes the feedback trend of the drip irrigation path in conjunction with the liquid state of the mixed fertilizer pump. The node scheduling order is adjusted according to the trend to obtain the control priority sequence. Based on the control priority sequence, the action execution module optimizes the action sequence of the partitioned solenoid valves, issues valve action commands in order, records the valve action response fed back by the controller, analyzes the time difference between the command sequence and the response process, and obtains the execution timing offset signal. Based on the execution timing offset signal, the steady-state monitoring module filters continuous periodic parameter changes, analyzes the stability state of the valve response process, records the control structure state when the response is consistently consistent, and supplements the scheduling record when response fluctuations are detected, thereby obtaining the control steady-state interval parameters.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, multi-channel signal standardization processing of feedback feature vector sets is used to achieve data fusion of root zone response and pipeline status. Node weight distribution information reflects differences in transport performance. The control priority sequence forms path sorting based on dynamic feedback trends. The execution time offset signal quantifies the consistency of action response. The control steady-state interval parameters support system adaptive optimization, forming a continuous adjustment and closed-loop response mechanism. This mechanism can flexibly respond to changes in crop root zone absorption and fluctuations in pipeline transport status, improving the coordination and operational reliability of the water and fertilizer control system. Attached Figure Description

[0015] Figure 1 This is a flowchart of the main steps of the present invention; Figure 2 This is a flowchart illustrating the process of obtaining the feedback feature vector set in this invention. Figure 3 This is a flowchart illustrating the process of obtaining node weight distribution information in this invention. Figure 4 This is a flowchart illustrating the process of obtaining the control priority sequence in this invention. Figure 5 This is a flowchart illustrating the acquisition of the timing offset signal in this invention; Figure 6 This is a flowchart illustrating the process of obtaining the steady-state range parameters in this invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0017] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0018] All user-related information involved in this invention (including but not limited to biometric information, identity verification information, behavioral data, device information, and other data that can be used for identity verification and personalized services) is collected and processed with the user's full knowledge and voluntary consent. The use of data is limited to purposes necessary for providing the technical services of this invention, and reasonable technical and management measures will be taken to ensure the security and confidentiality of users' personal information in terms of information protection and privacy.

[0019] For examples, please refer to Figure 1 This invention provides a technical solution: an AI-based method for regulating liquid fertilizer drip irrigation in agricultural planting, comprising the following steps: S1: Based on the operating data of liquid fertilizer storage tank and mixed fertilizer pump, analyze the changes in flow sensing signals inside the pipeline, detect the fluid flow rate change process in the drip irrigation main pipe, monitor the soil moisture response and root zone ion concentration signal response in the crop root zone through sensors, identify the amplitude change characteristics of each signal within the acquisition period, standardize the multi-channel signals, and obtain the feedback feature vector set. S2: Based on the feedback feature vector set, compare the differences of each parameter between different irrigation nodes, analyze the fluid transport status of each node in the drip irrigation main pipe, determine the node association based on the matching degree between the root zone feedback signal and the pipeline fluid status, calculate the transport difference between nodes, and obtain the node weight distribution information. S3: Based on the node weight distribution information, the root zone feedback response process of each drip irrigation zone is judged through the near-end strategy optimization algorithm. Combined with the fluid operation status of the fertilizer pump discharge stage, the feedback trend of each drip irrigation path is compared and analyzed. The node scheduling order is adjusted according to the feedback trend to obtain the control priority sequence. S4: Based on the control priority sequence, optimize the action sequence of the partition solenoid valves, issue valve action commands in sequence according to the scheduling order, record the valve action response fed back by the controller, analyze the time difference between the action command sequence and the actual response process, and obtain the execution timing offset signal; S5: Based on the execution timing offset signal, filter parameter changes within the continuous drip irrigation cycle, analyze the stability state of the valve response process, record the control structure state when the response is determined to be consistent, and supplement the scheduling record when the response process fluctuation is detected, so as to obtain the control steady-state range parameters.

[0020] The feedback feature vector set includes signal amplitude characteristics, signal response balance, and data consistency index within the period; the node weight distribution information includes node transmission efficiency coefficient, correlation strength parameter, and transmission difference distribution; the control priority sequence includes path priority number, feedback trend classification, and node sorting identifier; the execution timing offset signal includes valve action response time difference, trigger sequence matching degree, and execution offset parameter; the control steady-state interval parameter includes periodic stability discrimination parameter, interval switching identifier, and operating status number.

[0021] In S1, the change in flow sensing signal refers to the fluctuation of the electrical signal output by the flow sensor as the liquid fertilizer flows through the pipeline, caused by changes in the speed and flow rate of the fertilizer solution. It is mainly used to reflect the actual flow rate of the liquid. The fluid velocity change process refers to the dynamic change in the flow rate of the fertilizer solution within the drip irrigation main pipe during the irrigation cycle. This is usually acquired by flow or velocity sensors installed on the main pipe and used to determine the actual speed and stability of fertilizer delivery. Sensors generally refer to signal acquisition sensors installed in pipelines, soil, or crop root zones, mainly used to detect the real-time status of physical quantities such as flow rate, pressure, humidity, and ion concentration. Examples include flow sensors, pressure sensors, and soil moisture sensors. Sensors and conductivity sensors, etc.; Soil moisture response refers to the real-time changes in soil moisture in the crop root zone during drip irrigation, manifested as the rise and fall of soil water content, usually detected by soil moisture sensors; Root zone ion concentration signal response refers to the changes in the concentration of ion components (such as nitrogen, phosphorus, potassium, etc.) in the fertilizer solution in the root zone, acquired through ion-selective electrodes or conductivity sensors, reflecting the crop's nutrient absorption process; Amplitude variation characteristics refer to the summarization of the maximum, minimum, or average variation amplitude of various sensor signals in each sampling period, used to represent the overall fluctuation characteristics of the signal within that period; Standardization processing normalizes and numerically standardizes signals of different physical quantities, amplitudes, and dimensions.

[0022] In S2, irrigation nodes refer to locations in the drip irrigation system that can be independently controlled and have their data collected, including main pipes, branch pipes, and zone valves; fluid transport status refers to the flow of fertilizer solution within the pipeline, including parameters such as flow rate, velocity, and pressure, used to evaluate the transport capacity and status at each node; root zone feedback signals refer to the changes in parameters such as soil moisture and ion concentration in the crop root zone during irrigation, representing real-time feedback from the crop to water and fertilizer supply; pipeline fluid status refers to the current flow state of fertilizer solution in the pipeline, combined with parameters such as flow rate and pressure, reflecting the operating status of the transport system; node association determination refers to analyzing the synchronicity or correlation between root zone feedback signals and pipeline status to determine the control or influence relationships between nodes, forming logical connections between nodes; transport differences refer to the comparison results of fluid flow rate, velocity, and pressure parameters between different nodes, used to identify which path or node has problems such as uneven transport.

[0023] In S3, the root zone feedback response process refers to the dynamic response process of the crop root zone to irrigation water and fertilizer within the drip irrigation zone, including changes in humidity and ion concentration, reflecting the crop's absorption status; the fluid operation status refers to the flow and distribution of fertilizer solution in the output pipeline of the fertilizer pump during actual irrigation, which is acquired through relevant sensors; the drip irrigation path feedback trend refers to the direction and magnitude of changes in root zone feedback parameters under each drip irrigation path, which is an important basis for the system's automatic optimization control strategy; the node scheduling sequence refers to adjusting the execution order of each control node (such as zone valves) in the drip irrigation system based on the feedback of each node and the path performance, realizing dynamic priority management.

[0024] In S4, the action sequence refers to the order in which actuators such as zone solenoid valves are activated according to the control priority sequence; the valve action command refers to the control signal issued by the control system, which is used to drive the zone solenoid valves to open or close, thereby realizing the actual drip irrigation action; the execution controller refers to the equipment that receives the output action command and directly controls the valves, pumps and other actuators, and has the ability to collect action feedback; the action response refers to the feedback such as the time and status of the actual action after the valve or actuator receives the command, which is convenient for verifying whether the command is executed accurately.

[0025] In S5, parameter change refers to the dynamic change process of relevant control or feedback parameters (such as valve control response time, feedback signal amplitude, etc.) during the continuous operation of the drip irrigation system; stability state refers to the determination of whether the working process of valves and other actuators is consistent and reliable through multi-cycle response and parameter change trend analysis; continuous response consistency means that the response time and feedback performance of the actuators remain basically constant during the continuous irrigation cycle, reflecting the stability of the control system; control structure state refers to the parameter combination and control logic state formed by the drip irrigation control system in a certain stable operating stage; supplementary scheduling record means that when fluctuations in the response of the actuators are detected, the system will save and mark the abnormal data of the current cycle for subsequent adaptive correction.

[0026] Please see Figure 2 The specific steps for obtaining the feedback feature vector set are as follows: S111: Based on the operating data of liquid fertilizer storage tank and mixed fertilizer pump, analyze the changes in the output signal of the flow sensor as the fertilizer solution flows in the pipeline, compare the data sequence synchronously collected by the pressure sensor, calculate the continuous distribution of the flow velocity at each monitoring point over time, screen the time synchronization segments of soil moisture response and ion concentration response in the crop root zone, determine the data mapping of each channel signal at the same time, and obtain multi-channel sensor response data. Based on the operational data of the liquid fertilizer storage tank and the mixing pump, a flow sensor was installed near the outlet of the main delivery pipeline. The sensor's output electrical signal change value was collected 10 times per second, and the electrical signal was converted into digital flow velocity data via an analog-to-digital converter. The flow rate of the fertilizer solution in the pipeline was recorded per second. Simultaneously, a pressure sensor was installed adjacent to the flow sensor, and pressure data in the pipeline was collected synchronously at the same sampling frequency. The flow velocity and pressure data were synchronized and time-series matched to calculate the continuous distribution of flow velocity at each monitoring point over time. When the flow velocity changed by more than 0.5 liters per minute and the corresponding pressure changed by more than 10 kPa within 10 consecutive seconds, the corresponding time interval was marked as a flow velocity abrupt change interval. Additionally, a humidity sensor and an ion concentration detection unit were deployed in the soil of the crop root zone. The ion concentration was determined based on the soil solution electrical conductivity collected by the conductivity sensor. Conductivity data, equivalent ion concentration parameters calculated through a preset conversion relationship, and soil moisture content and ion concentration data collected every 5 seconds by a humidity sensor and ion concentration detection unit. Taking the start time of each irrigation as the starting point of a unified timeline, the data within a 10-second range before and after the flow velocity change segment are compared and analyzed. Time segments with a soil moisture content increase greater than 3% and an ion concentration increase greater than a preset threshold are selected to determine the time synchronization segment of soil moisture response and ion concentration response in the crop root zone. The data of each sensor channel within the time synchronization segment are compared to determine whether there is a data mapping relationship of continuous response trend of flow velocity, pressure, soil moisture content and ion concentration within the same time range. When the data of each channel shows a continuous upward trend within the time range, the corresponding time period and its value combination are retained, and all time period data that meet the conditions are summarized to form multi-channel sensor response data.

[0027] S112: Based on multi-channel sensor response data, analyze the fluctuation range of each channel signal within the complete cycle, calculate the amplitude change of each sampling point, identify continuous data segments within the cycle boundary, compare the dynamic response between different signals, and obtain the channel fluctuation amplitude set. For four channels—flow rate, pressure, humidity, and ion concentration—the differences between their maximum and minimum values ​​throughout the entire irrigation cycle are statistically analyzed to obtain the fluctuation range of each channel. For example, if the flow rate channel has a maximum of 3.1 liters / minute and a minimum of 1.7 liters / minute, the fluctuation range is 1.4 liters / minute. Similarly, the humidity channel has a maximum of 29% and a minimum of 21%, with a fluctuation range of 8%. This process is repeated to calculate the set of amplitude changes for the four channels. Then, using a 5-minute time window, all values ​​within that time period are extracted from the data of each channel. The maximum and minimum differences within the window are calculated. If this difference reaches more than 50% of the total amplitude of that channel, the segment is defined as a significant fluctuation segment. For example, if the total amplitude of the humidity channel is 8%, the value increases from 22% to 27% within a 5-minute window. A fluctuation of 5% or more, accounting for more than 60%, is considered an effective fluctuation segment. A 5-minute buffer is established at the beginning and end of each cycle, extending forward and backward respectively. Within this buffer, if any channel experiences a fluctuation exceeding 30% of its own amplitude, it is considered a cycle boundary response segment. For example, if the pressure channel amplitude is 20 kPa, and the initial fluctuation is 7 kPa, this segment is retained as a boundary response. The start and end times of each channel fluctuation segment are then compared with the data number to determine if different channels change in the same direction within the same time period. If multiple channels show an upward or downward trend within 10 seconds with a change ratio exceeding 30%, this time period is marked as a multi-channel dynamic response segment. All effective fluctuation segments and dynamic response segments are categorized and summarized into a channel fluctuation amplitude set.

[0028] S113: Based on the set of channel fluctuation amplitudes, analyze the overall distribution of amplitude changes in all channels, calculate the feature vectors under a unified scale, determine the data balance in the periodic response, optimize the expression format of cross-channel changes, and obtain the set of feedback feature vectors; The fluctuation amplitude of each channel is uniformly converted to a value between 0 and 1. For example, a flow rate channel fluctuation of 1.4 L / min has a maximum value of 1.4 and a minimum value of 0.6, which is converted to a value of 1.0. A humidity channel fluctuation of 8% has a maximum value of 10% and a minimum value of 2%, which is converted to 0.75. The amplitudes of the four channels are converted to dimensionless proportional values, resulting in values ​​such as flow rate 1.0, pressure 0.65, humidity 0.75, and ion concentration 0.5. Then, with each minute as a sub-cycle, the normalized values ​​of the four channels are extracted within each sub-cycle. The average difference between each pairwise value is calculated. If the average difference is less than 0.2, the channel response within that minute is considered consistent and is recorded as an equilibrium segment. If, at the 8th minute, the values ​​of the four channels are 0.7, 0.6, 0.8, and 0.75 respectively, with an average difference of 0.1, then the condition is met. Record the time number and channel status, and then splice several consecutive equilibrium segments to form a stable period segment. After that, splice the amplitude values ​​of each channel in each stable period segment in the channel order to form a feedback vector. For example, if the order is flow rate-pressure-humidity-concentration, the values ​​are 0.7, 0.6, 0.8, and 0.75 respectively, forming a vector [0.7, 0.6, 0.8, 0.75]. Each feedback vector represents the complete response data combination of the crop root zone and pipeline channel under one irrigation cycle. Collect all feedback vectors under multiple consecutive cycles and output as a set of feedback feature vectors.

[0029] Please see Figure 3 The specific steps for obtaining node weight distribution information are as follows: S211: Based on the feedback feature vector set, analyze the corresponding parameters of each irrigation node, compare the amplitude feature differences between nodes in the same irrigation cycle, determine the signal response changes of each node in the same time period, calculate the spatial distribution range of node parameters, and obtain the irrigation node parameter difference sequence. First, each irrigation node in the drip irrigation system is numbered, such as N0 for the main pipeline and N1 to N10 for each branch node. The channel response parameters recorded in the feedback feature vector set are then mapped to nodes. The amplitude features of flow rate, pressure, humidity, and ion concentration for each node within the current irrigation cycle are extracted. The difference between the maximum and minimum values ​​in the response data of each node is calculated to construct the amplitude feature vector for that node. For example, node N3 might have a flow rate fluctuation of 1.3 L / min, a pressure fluctuation of 15 kPa, a humidity fluctuation of 6%, and an ion concentration fluctuation of 0.2 mSiemens / cm² within this cycle. These four values ​​are combined to form the node response vector. Then, the amplitude differences between any two nodes in the same cycle are compared pairwise. If the difference between any channel in two nodes exceeds 40% of the average amplitude of that channel across the entire system, it is recorded as a significantly different node pair. For example, if the average humidity amplitude is 5%, and N3 is 6% and N7 is 3%, the difference is 3%, accounting for 60%, which is considered a significant difference. A comparison matrix is ​​then constructed among all nodes. The comparison results between all nodes are recorded sequentially to determine the shape changes of the channel response curves of each node within the same time period. When the flow or pressure fluctuations of a certain node are concentrated in the early part of the irrigation cycle, while other nodes are concentrated in the middle and late parts, the response of that node has a time drift. It is necessary to mark its response curve by comparing the response timestamps, divide the high response interval of each node, and then compare the position number of each node in the actual field layout according to its spatial location. Calculate the concentration of the amplitude change range in spatial distribution. For example, if N3, N4, and N5 are in the same group in physical layout and have similar amplitude characteristics, they are classified into the same segment. At the same time, compare the response value difference between this area and other areas, and calculate the numerical span between the maximum and minimum response amplitude in the spatial distribution range. If the span is greater than the set threshold, such as the span exceeding 20 kPa in the pressure channel, it is recorded as a high difference area. Establish a record table for the amplitude comparison results and spatial location correspondence of all nodes, arrange them in order of node number, and output the irrigation node parameter difference sequence.

[0030] S212: Based on the difference sequence of irrigation node parameters, analyze the fluid velocity changes of each node, compare the fluid continuous transmission state between nodes, determine the temporal response consistency between the root zone feedback signal and the main fluid state, identify nodes with correlation within the time period, adjust the synchronization structure of the response data, and obtain the node synchronization response correlation data. The fluid velocity sequence of each node within the same irrigation cycle is retrieved, and the velocity values ​​between adjacent nodes are extracted. The velocity change between every two consecutive nodes is calculated. For example, if the average velocity of node N3 is 2.5 L / min and that of N4 is 2.1 L / min, the change is 0.4 L / min. This calculation process is performed on all consecutive nodes, and node pairs with velocity changes greater than 0.6 L / min are selected as discontinuous state points. The shape of the velocity curve of each node's path within the irrigation cycle is then compared. If the curve trends are inconsistent (e.g., one node shows a single peak change while the other shows a step-like increase), the path is considered an unstable transmission path. Furthermore, the humidity and ion concentration change curves in the root zone feedback signal are extracted to establish a correlation with the fluid velocity change sequence in the main pipeline. The time correspondence is determined by comparing the time difference between the root zone response time point and the fluid velocity change time point in 5-second intervals. If the time difference is less than 10 seconds and occurs more than 3 times consecutively, the timing response is considered consistent. Related node pairs are constructed according to the path pair method. For example, if N3 and N4 simultaneously meet the conditions of stable flow velocity and synchronized root zone feedback, the path is marked as a related path. All related path data are then uniformly organized, and the timestamps of the response data are restructured to align their response values ​​on a unified timeline. For example, the response values ​​of N3 and N4 are aligned to the same 10-second period. The data structure in the time dimension is rebuilt in this way, and all adjusted node pair response data are summarized, outputting the node synchronization response related data.

[0031] S213: Based on the node synchronous response correlation data, calculate the offset parameters of the transport state between each node, analyze the coupling strength between the node feedback signal and the fluid transport characteristics, determine the control capability of the master control node, optimize the mapping structure between node number and transport parameters, and obtain the node weight distribution information. First, the offset of fluid transport status is calculated between each pair of nodes. The offset value is calculated based on the difference between data such as flow rate and pressure recorded in the same sampling period. For example, if the flow rate difference between nodes N3 and N4 is 0.4 L / min and the pressure difference is 9 kPa at a certain moment, the value is recorded as the transport offset parameter of that node pair. After continuously collecting the offset values ​​for multiple time periods, the standard deviation is calculated. When the standard deviation is greater than a set threshold, such as the flow rate offset standard deviation exceeding 0.6 L / min, it is judged that the transport of that node is unstable. Then, the maximum fluctuation of the root region feedback value is extracted from each node and compared with the synchronization degree between the fluid transport data. The time interval in which the change trends of any two are similar is calculated as the proportion of the entire period. The proportion of nodes in the pipeline network is used to determine the degree of coupling. If the proportion is higher than 70%, the node is considered a highly coupled node. The control influence range of each node is statistically analyzed according to its position in the pipeline network. For example, if node N3 is the starting partition node, and N4 to N6 are its downstream nodes, and all three are highly coupled, then N3 is marked as the master control node. The sum of the number of its control nodes and the feedback intensity is used as its control capability parameter. The numbering system is further re-established according to the node number, so that the master control node number comes first, and the downstream nodes are arranged in order of their response intensity to the master control node. A mapping relationship is established between the number of each node and its corresponding transmission parameters. Key data such as response amplitude, time delay and synchronization ratio in the mapping are recorded to obtain the node weight distribution information.

[0032] Please see Figure 4 The specific steps for obtaining the priority sequence are as follows: S311: Based on node weight distribution information, the operation data of drip irrigation nodes is analyzed through a near-end strategy optimization algorithm. The combined processing of root zone humidity response and nutrient concentration change is performed to determine the dynamic synchronicity of crop root zone water change and ion concentration fluctuation between drip irrigation paths. The fluid transport status of the irrigation path is synchronously detected to obtain the path feedback trend set. For each node, the transmission efficiency, correlation strength, and transmission difference within the same sampling period are numerically read and arranged sequentially. For example, if the sampling period is set to 10 seconds, the corresponding values ​​for a certain node within a certain period are 0.82, 0.65, and 0.18, respectively. The three values ​​are proportionally processed: transmission efficiency is multiplied by 0.5 to get 0.41, correlation strength is multiplied by 0.3 to get 0.195, and transmission difference is subtracted from 1 to get 0.82, then multiplied by 0.2 to get 0.164. The three results are added together to get 0.769, which is taken as the current state value of the node. Then, the humidity change data of the corresponding path of the node within the same period is read. For example, humidity changing from 18 to 26 results in a change of 8, and conductivity changing from 1.2 to 1.8 results in a change of 0.6. The two are divided to get approximately 13. This value is divided into three intervals: 0 to 5, 5 to 10, and greater than 10. If the value falls within the range greater than 10, then the difference between the humidity change and conductivity change within three consecutive cycles is calculated. For example, the humidity change is 3, 2, 1, and the conductivity change is 0.2, 0.1, 0.3. For each cycle, a sign consistency judgment is performed and the difference is calculated. For example, if the difference in the first cycle is 2.8, which is greater than 2, it is recorded as unsatisfied; if the difference in the second cycle is 1.9, which is less than or equal to 2, it is recorded as satisfied; and if the difference in the third cycle is 0.7, it is recorded as satisfied. The number of satisfied cycles is 2, and the total number of satisfied cycles is 3, resulting in a ratio of approximately 0.67. The ratio is divided into three intervals: less than 0.4, 0.4 to 0.7, and greater than 0.7. The path falls within the 0.4 to 0.7 interval. At the same time, the flow rate change ratio of 0.12 and the pressure change ratio of 0.1 in this cycle are read. The difference between the two is 0.02, which is less than 0.1, and it is recorded as a consistent state. The synchronization level and the consistent state are combined, and the path feedback trend set is obtained after executing the path one by one.

[0033] S312: Based on the path feedback trend set, compare it with the fluid transport characteristics during the operation of the fertilizer pump, calculate the fluctuation amplitude of each path feedback and the normalized result of the pipeline pressure response, obtain the feedback trend intensity factor, determine the strength relationship of the feedback trends between each path, and obtain the node control ranking information. The trend indicator corresponding to each path is matched with the fertilizer pump operating cycle data cycle by cycle. For example, the fertilizer pump operating time is set to 60 seconds and divided into 6 cycles, each cycle being 10 seconds. Within a certain cycle, the path pressure change range is read, for example, from 0.18 to 0.22, resulting in a change of 0.04. Simultaneously, the corresponding cycle pressure change of the fertilizer pump is read as 0.05. The path pressure change is divided by the fertilizer pump change, resulting in 0.8. This result is divided into three levels: 0 to 0.5, 0.5 to 0.8, and greater than 0.8. This value falls within the 0.5 to 0.8 range. Then, the humidity change is read as 6 and the conductivity change as 0.5 within the same cycle. Add the two together to get 6.5. Then multiply 6.5 by 0.8 to get 5.2. Divide the result into three intervals: 0 to 3, 3 to 6, and greater than 6. The value is in the interval 3 to 6. Repeat the above calculation for all paths. For example, if another path calculates to 6.8, it falls into the interval greater than 6. Then compare the results of all paths one by one. For example, 6.8 minus 5.2 equals 1.6, which is greater than 1. Then mark the former as having priority. If the difference is less than 1, mark them as being of the same level. Organize all comparison results into an order relationship. For example, path 2 has priority over path 1, which has priority over path 3. Then convert the order into a number form and assign the values ​​1, 2, and 3 respectively to obtain the node control sorting information.

[0034] S313: Based on the node control and sorting information, prioritize all drip irrigation paths, mark the feedback trend level and the identification number of the associated node, optimize the path arrangement order, and obtain the control priority sequence. All paths are arranged in ascending order of their sort numbers. For example, path 2 is numbered 1, path 1 is numbered 2, and path 3 is numbered 3. Simultaneously, the trend level corresponding to each path is read and numerically processed: weak level is assigned a value of 1, medium level is assigned a value of 2, and strong level is assigned a value of 3. For example, path 2 is assigned a value of 3, path 1 is assigned a value of 2, and path 3 is assigned a value of 1. During priority reordering, paths are first sorted by number, and then paths with the same number are sorted again by trend level from largest to smallest. For example, if two paths are both numbered 2, the one with trend level 3 is placed first. The sorted paths are then reassigned accordingly. Priority numbers are assigned, for example, the first path is numbered 1, the second is numbered 2, and the third is numbered 3. At the same time, the node number corresponding to each path is read. For example, if the node number corresponding to path 2 is 5, then the record is 1 corresponding to 5. Then, the continuity of the entire sequence is checked, and the numbers are compared item by item to see if they increase sequentially. For example, if 1, 3, and 4 appear, 3 is adjusted to 2 and 4 is adjusted to 3 to make the numbers continuous. Then, the priority number, trend level value, and node number of each path are combined into a group of records, such as 1, 3, and 5. After all paths are processed, they are arranged and output in priority order to obtain the control priority sequence.

[0035] Please see Figure 5The specific steps for obtaining the timing offset signal are as follows: S411: Based on the control priority sequence, control commands are sent to the solenoid valve in sequence to optimize the action signal issuance process. Time synchronization is used to record the issuance time of each command in a unified manner, and the scheduling order of each command is determined to obtain the valve command scheduling time sequence set. The path priority number is mapped one-to-one with the control number of the partition solenoid valve. The controller's built-in instruction cache module arranges the control instruction queue according to priority. Each control instruction includes fields such as solenoid valve number, target action status, and trigger delay time. The base time step is set to 100 milliseconds. Start signals are sent to the corresponding solenoid valves sequentially according to their numbers. For example, if the priority sequence is [P3, P1, P4, P2], the corresponding sending order is solenoid valves V3-V1-V4-V2. Before issuing the instruction, the controller synchronously calls the local time base module to generate a current timestamp, recording the specific millisecond moment of each instruction transmission. The target solenoid valve number and instruction sequence number are marked in the record, forming a preliminary scheduling record. After all control instructions have been issued, the records for each solenoid valve are retrieved again. The command issuance time is used to generate a scheduling timeline, sorted by time. The time intervals of any adjacent commands are compared. If the interval is less than the preset minimum safety interval (e.g., 200 milliseconds), the issuance time of the next command is adjusted to ensure that the interval between all commands is not less than this baseline value. The updated schedule is then executed to send control signals sequentially. If a command is not responded to within the set time period or the response delay exceeds 500 milliseconds, the anomaly is recorded and its position in the original scheduling order is marked. At the same time, information such as the solenoid valve number, the original sending time, and the actual response time are retained. The time distribution of all commands is extracted from the entire scheduling record to generate structured scheduling time sequence data containing fields such as scheduling order, solenoid valve number, issuance time, and response delay, which serves as the valve command scheduling time sequence set.

[0036] S412: Based on the valve command scheduling timing set, synchronously collect the response time of each solenoid valve action fed back by the controller, calculate the time interval between the action response and the command issuance, using the formula: ; Obtain the response time offset of each solenoid valve to obtain a response offset vector set, where, This represents the response time offset of the z-th solenoid valve. This indicates the time of the action response of the z-th solenoid valve, obtained using a time synchronization method. This indicates the time when the z-th solenoid valve receives the command, recorded using a unified time base. This represents the distance normalization coefficient of the fluid path containing the z-th solenoid valve, obtained by standardizing the actual path length. This indicates the current command execution sequence number for the z-th solenoid valve, reflecting the priority of command processing. This indicates the length of the command buffer queue in the partition containing the z-th solenoid valve, reflecting the number of instructions to be processed. This indicates the current fluid pressure level at the position of the z-th solenoid valve, using pressure parameters measured by a sensor. Response time offset refers to the time interval between the actual time point when the control system issues an action command to the z-th solenoid valve and the actual time point when the command is issued, after fluid path distance normalization and weighting by parameters such as command sequence, buffer queue, and current pressure. Response time offset comprehensively reflects whether the solenoid valve's action response is synchronized with the system's scheduling expectations, and can quantify the time offset phenomenon caused by multiple factors such as physical path, execution sequence, and buffer pressure during valve action. Synchronously retrieve the action response data of each solenoid valve in the controller feedback channel to obtain the actual occurrence time of each solenoid valve's action. And record the time when the corresponding control command is issued. The time difference between the two is calculated as the basic response interval. Then, the physical length of the path where the corresponding solenoid valve is located, the control sequence number, the command buffer queue length, and the real-time pressure value of the current node pipeline are collected and used as parameters. , , , The calculation of response time offset involves normalizing parameters of different dimensions. The normalization method uses interval scaling, which scales the parameters to the [0, 1] interval based on their maximum and minimum values ​​defined within the control system. For example, the parameter settings and normalized data for a solenoid valve are as follows: Command issuance time: (System master clock record); Response time: (Recorded during controller state transition); Original physical distance of path: 36m, normalized. (Maximum path length is set to 50m); Control command sequence number: (No normalization, actual number used); Command cache length original value: 5 entries (taken from the current state of the cache queue), no normalization; Real-time pressure original value: 400 (Data collected from node pressure sensor), after normalization (The maximum pressure is set to 500 kPa); Substitute into the formula to calculate the time difference: ; Calculate the numerator: ; Calculate the denominator: ; Calculation results: ; The interval is divided as follows: When Classified as a synchronous response segment, it indicates that the solenoid valve response is highly consistent and the control path operates stably; when This is classified as a medium offset segment, indicating that the response has a certain lag but is still within the acceptable control range, and subsequent response comparison and order optimization can be performed; when This value is classified as a significant offset segment, indicating an abnormal delay in the solenoid valve's response. It needs to be included in the control record and used as a basis for scheduling adjustments. Therefore, this offset result value... In The interval segment belongs to the medium offset segment. This offset value serves as an input in the response offset vector set and can be called into the subsequent execution timing offset signal construction stage. It will be sorted, graded, and logically judged with other solenoid valve offset values, and used to form the basis for judging whether the timing of the control action is consistent. The magnitude of this value directly determines its sorting weight and logical position in the overall offset signal.

[0037] S413: Based on the response offset vector set, filter all signals with sequential offsets, determine the temporal sequence relationship of each group of signals, and archive the offset time and sequence characteristics to obtain the execution timing offset signal; The control command records in the scheduling sequence set are read one by one. The command issuance time is compared with the actual response time of the solenoid valve feedback action. The time difference between the two is calculated as the response offset value. For example, if the control command for solenoid valve V3 is issued at 10:00:00.300 seconds and the actual response time is 10:00:00.700 seconds, the offset is 400 milliseconds. The response offset values ​​of all solenoid valves within a scheduling cycle are statistically analyzed, and records with offset times greater than 300 milliseconds are filtered out as signals of significant timing offset. These signals are then analyzed again to determine their relative position in the scheduling sequence to determine if there is any abnormal order. If the response time of a lower priority solenoid valve is more than 200 milliseconds earlier than that of a higher priority solenoid valve, it is determined that there is a reversed order behavior. For example, V4 V1 is priority 1, V2 is priority 4. If V2 responds at 10:00:01.100 seconds and V4 responds at 10:00:01.400 seconds, with a sequence offset of 300 milliseconds, then this record is classified as a sequence anomaly. Subsequently, all offset values ​​are sorted according to the response time, and the offset time interval distribution of each group of signals is established. The offset level is divided into 100-millisecond units, with 0–200 milliseconds classified as normal offset, 200–500 milliseconds as moderate offset, and more than 500 milliseconds as severe offset. Information such as the level of each signal, the specific time difference, and the control number is recorded. At the same time, a sequence offset event table is established to summarize all solenoid valve groups with misaligned responses, marking their original order, response order, and offset time, and generating an execution timing offset signal containing the offset time level, response order relationship, and control number structure.

[0038] Please see Figure 6 The specific steps for obtaining the parameters for the steady-state range are as follows: S511: Based on the execution timing offset signal, analyze the response sequence of each valve in the continuous drip irrigation cycle, compare the distribution of the start and end nodes of the valve response in different cycles, determine whether the unfolding order of the response process in each cycle is stable, identify response segments with similar unfolding trajectories, and obtain cycle response stability data. The response records of each solenoid valve in a continuous drip irrigation cycle are extracted sequentially. Within each cycle, the timestamp of the issued command is compared with the actual response time, and the response start time and action completion time are marked. A response sequence list is generated with a time granularity of 10 milliseconds. The response sequence of each valve in each cycle is archived by number. For example, the response start time of solenoid valve V1 in cycle T1 is 10:00:00.200 seconds and the end time is 10:00:00.600 seconds, while in cycle T2 they are 10:30:00.250 seconds and 10:30:00.650 seconds respectively. The response time periods of this valve in all cycles are recorded sequentially. Then, the data is aligned by cycle to construct a cycle comparison table to compare the same solenoid valve in different cycles. The difference between the start and end times of the valve's response is determined. If the start time difference is less than 100 milliseconds and the end time difference is less than 120 milliseconds, the start and end node distributions are considered consistent. Otherwise, they are marked as changing nodes. The response stability percentage of each solenoid valve in all cycles is calculated. When the stable distribution percentage exceeds 80%, it is considered a stable valve. The response order of all valves in the same cycle is listed in sequence to construct a response trajectory sequence. The trajectory is then compared with the previous cycle. If the order of the first 5 valves in the sequence is completely consistent, the trajectory segment is marked as a similar unfolded trajectory. If at least two cycles in more than 3 cycles maintain the same trajectory, it is summarized as a stable unfolded segment. All response segments that meet the trajectory stability requirements are numbered and output to form the periodic response stability data.

[0039] S512: Based on periodic response stability data, analyze the response of drip irrigation paths, compare the response order of each path in a continuous period, determine the consistency of the path response order in different periods, identify path segments with continuous response order, and perform hierarchical labeling on segments where the response order changes to obtain drip irrigation path mapping results. The response records of the solenoid valves associated with each drip irrigation path are extracted within consecutive cycles. The response numbers of each path within each cycle are sorted and numbered chronologically to construct a path response sequence. For example, if the response order of path P1 in cycle T1 is 2nd, in T2 it is 2nd, and in T3 it is 3rd, then the response positions of this path in the three cycles are recorded as [2, 2, 3]. The sequence is then evaluated using sequence difference. If the sequence difference between any two consecutive cycles is less than or equal to 1, and there is at most one change, then the path is marked as a sequence-maintaining path; otherwise, it is considered a switching path segment. The response positions of paths P1 and P3 in the three consecutive cycles are [1, 1, 1] and [4, 3, 5], respectively. P1 is determined to be continuously maintained. P3 is recorded as a switching segment because it has two jump differences greater than 1 position. Then, the switching path is classified according to the direction of change of the response position. If it changes from low position to high position, it is marked as sequential shift type, otherwise it is early response type. Then, according to the position of the stable path in the cycle, the entire path structure is divided into multiple segments. For example, the first 3 paths are marked as high priority segments, the middle 3 as middle segments, and the last as low priority segments. Paths with consistent response order in the same segment are grouped together and their path number and sequence identifier are marked. The output contains a dataset containing path number, sequence sequence in the cycle, whether it is stable, and switching flag, which constitutes the drip irrigation path mapping result.

[0040] S513: Based on the drip irrigation path mapping results, analyze the scheduling and control characteristics within each segment, determine the continuity of the response structure within each path segment, identify scheduling parameters that exhibit consistency in continuous segments, optimize the numbering and scheduling relationship of each segment, and obtain the steady-state interval parameters for regulation. Extract the path number and key parameters such as action time, response delay, and control signal strength within each segment and its control cycle. Aggregate the scheduling control feature vector of each segment by path number. Compare the time interval between the control signal and response feedback for all paths within each segment. If the maximum time difference between the control signal arrival and response feedback for a path does not exceed 100 milliseconds across all cycles, the path scheduling structure is marked as continuous. If this phenomenon accounts for more than 80% of the paths in the entire segment, the segment is determined to be a continuous response segment. Then, average the response delay time within the segment and extract the average value. The path whose delay is closest to the overall mean of the segment is taken as the representative path of the segment. Its number and its corresponding control parameter set are recorded. At the same time, the numbers of all paths in the segment are extracted, and the sorting relationship based on the representative path is reconstructed. The sorted path numbers are reassigned to the segment numbers. For example, if the continuous segment number is Z1 and the corresponding path number is [P3, P5, P6], and the response delay is sorted from smallest to largest as P5, P6, P3, then the segment number mapping relationship is Z1-P5, Z1-P6, Z1-P3. Finally, the control characteristic parameters, sorting structure, and number mapping of all segments are integrated, and the output is the control steady-state interval parameter.

[0041] An AI-based drip irrigation control system for liquid fertilizer in agricultural planting, comprising: The feedback acquisition module is based on the liquid fertilizer storage tank, analyzes the changes in flow signals in the pipeline, detects the changes in the flow rate of the drip irrigation main pipe, monitors the response of soil moisture and ion concentration in the crop root zone, identifies the periodic signal amplitude characteristics, and obtains the feedback feature vector set. The weight analysis module compares the performance of each parameter across different irrigation nodes based on the feedback feature vector set, analyzes the delivery status of the drip irrigation main node, determines the node association based on the matching degree between the root zone feedback signal and the pipeline status, calculates the delivery difference, and obtains the node weight distribution information. The priority determination module judges the root zone feedback response of the drip irrigation zone based on the node weight distribution information, and analyzes the feedback trend of the drip irrigation path in conjunction with the liquid state of the mixed fertilizer pump. Based on the trend, the node scheduling order is adjusted to obtain the control priority sequence. The action execution module optimizes the action sequence of the partitioned solenoid valves based on the control priority sequence, issues valve action commands in order, records the valve action response fed back by the controller, analyzes the time difference between the command sequence and the response process, and obtains the execution timing offset signal. The steady-state monitoring module filters continuous periodic parameter changes based on the execution timing offset signal, analyzes the stability state of the valve response process, records the control structure state when the response is consistently consistent, and supplements the scheduling record when response fluctuations are detected, thus obtaining the control steady-state range parameters.

[0042] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. An AI-based liquid fertilizer drip irrigation control method for agricultural planting, characterized in that, Includes the following steps: S1: Based on the operating data of liquid fertilizer storage tank and mixed fertilizer pump, analyze the changes in flow signal in the pipeline, detect the changes in the flow rate of the drip irrigation main pipe, monitor the soil moisture and ion concentration response in the crop root zone through sensors, identify the periodic signal amplitude characteristics, and obtain the feedback feature vector set; S2: Based on the feedback feature vector set, compare the performance of each parameter among different irrigation nodes, analyze the delivery status of the drip irrigation main node, determine the node association based on the matching degree between the root zone feedback signal and the pipeline status, calculate the delivery difference, and obtain the node weight distribution information. S3: Based on the node weight distribution information, the root zone feedback response of the drip irrigation zone is judged by the near-end strategy optimization algorithm. Combined with the liquid state of the mixed fertilizer pump, the feedback trend of the drip irrigation path is analyzed by comparison. The node scheduling order is adjusted according to the trend to obtain the control priority sequence. S4: Based on the control priority sequence, optimize the action sequence of the partitioned solenoid valves, issue valve action commands in order, record the valve action response fed back by the controller, analyze the time difference between the command sequence and the response process, and obtain the execution timing offset signal; S5: Based on the execution timing offset signal, filter continuous periodic parameter changes, analyze the stability state of the valve response process, record the control structure state when the response is consistently consistent, and supplement the scheduling record when response fluctuations are detected, so as to obtain the control steady-state interval parameters. 2.The AI-based liquid fertilizer drip irrigation regulation method for agricultural planting of claim 1, wherein, The feedback feature vector set includes signal amplitude features, signal response balance, and data consistency index within the period; the node weight distribution information includes node transmission efficiency coefficient, correlation strength parameter, and transmission difference distribution; the control priority sequence includes path priority number, feedback trend classification, and node sorting identifier; the execution timing offset signal includes valve action response time difference, trigger sequence matching degree, and execution offset parameter; the control steady-state interval parameter includes periodic stability discrimination parameter, interval switching identifier, and operating status number. 3.The AI-based liquid fertilizer drip irrigation regulation method for agricultural planting of claim 1, wherein, The specific steps for obtaining the feedback feature vector set are as follows: S111: Based on the operating data of liquid fertilizer storage tank and mixed fertilizer pump, analyze the changes in the output signal of the flow sensor as the fertilizer solution flows in the pipeline, compare the data sequence synchronously collected by the pressure sensor, calculate the continuous distribution of the flow velocity at each monitoring point over time, screen the time synchronization segments of soil moisture response and ion concentration response in the crop root zone, determine the data mapping of each channel signal at the same time, and obtain multi-channel sensor response data. S112: Based on the multi-channel sensing response data, analyze the fluctuation range of each channel signal within the complete cycle, calculate the amplitude change of each sampling point, identify continuous data segments within the cycle boundary, compare the dynamic responses between different signals, and obtain the channel fluctuation amplitude set. S113: Based on the set of channel fluctuation amplitudes, analyze the overall distribution of amplitude changes in all channels, calculate the feature vectors under a unified scale, determine the data balance in the periodic response, optimize the expression format of cross-channel changes, and obtain the feedback feature vector set. 4.The AI-based liquid fertilizer drip irrigation regulation method for agricultural planting of claim 1, wherein, The specific steps for obtaining the node weight distribution information are as follows: S211: Based on the feedback feature vector set, analyze the corresponding parameters of each irrigation node, compare the amplitude feature differences between nodes in the same irrigation cycle, determine the signal response changes of each node in the same time period, calculate the spatial distribution range of node parameters, and obtain the irrigation node parameter difference sequence. S212: Based on the difference sequence of irrigation node parameters, analyze the fluid velocity changes of each node, compare the fluid continuous transmission state between nodes, determine the temporal response consistency between the root zone feedback signal and the main fluid state, identify nodes with correlation within the time period, adjust the synchronization structure of the response data, and obtain node synchronization response correlation data. S213: Based on the node synchronous response association data, calculate the offset parameters of the transport state between each node, analyze the coupling strength between the node feedback signal and the fluid transport characteristics, determine the control capability of the master control node, optimize the mapping structure between node number and transport parameters, and obtain node weight distribution information. 5.The AI-based liquid fertilizer drip irrigation regulation method for agricultural planting of claim 1, wherein, The specific steps for obtaining the regulation priority sequence are as follows: S311: Based on the node weight distribution information, the operation data of the drip irrigation nodes are analyzed through the near-end strategy optimization algorithm. The humidity response and nutrient concentration changes in the root zone of the irrigation area are combined and processed to determine the dynamic synchronicity of crop root zone water changes and ion concentration fluctuations between drip irrigation paths. The fluid transport status of the irrigation path is synchronously detected to obtain the path feedback trend set. S312: Based on the path feedback trend set, compare it with the fluid transport characteristics during the operation of the fertilizer pump, calculate the normalized result of the fluctuation amplitude of each path feedback and the pipeline pressure response, obtain the feedback trend intensity factor, determine the strength relationship of the feedback trends between each path, and obtain the node control sorting information. S313: Based on the node control and sorting information, prioritize all drip irrigation paths, mark the feedback trend level and associated node identification number, optimize the path arrangement order, and obtain the control priority sequence. 6.The AI-based liquid fertilizer drip irrigation regulation method for agricultural planting of claim 1, wherein, The specific steps for obtaining the execution timing offset signal are as follows: S411: Based on the control priority sequence, control commands are sent to the solenoid valve in sequence to optimize the action signal issuance process. The issuance time of each command is uniformly recorded using time synchronization, the scheduling order of each command is determined, and a valve command scheduling time sequence set is obtained. S412: Based on the valve command scheduling timing set, synchronously collect the action response time of each solenoid valve fed back by the controller, calculate the time interval between the action response and the command issuance, obtain the response time offset of each solenoid valve, and obtain the response offset vector set. S413: Based on the response offset vector set, filter all signals with sequential offsets, determine the temporal sequence relationship of each group of signals, and archive the offset time and sequence characteristics to obtain the execution timing offset signal. 7.The AI-based liquid fertilizer drip irrigation regulation method for agricultural planting of claim 1, wherein, The specific steps for obtaining the parameters of the control steady-state range are as follows: S511: Based on the execution timing offset signal, analyze the response sequence of each valve in the continuous drip irrigation cycle, compare the distribution of the start and end nodes of the valve response in different cycles, determine whether the unfolding order of the response process in each cycle is stable, identify response segments with similar unfolding trajectories, and obtain cycle response stability data. S512: Based on the periodic response stability data, analyze the drip irrigation path response, compare the response order of each path in a continuous period, determine the consistency of the path response order in different periods, identify the path segments whose response order remains continuous, and perform hierarchical identification on the segments where the response order changes, to obtain the drip irrigation path mapping result. S513: Based on the drip irrigation path mapping results, analyze the scheduling and control characteristics within each segment, determine the continuity of the response structure within each path segment, identify scheduling parameters that exhibit consistency in continuous segments, optimize the numbering and scheduling relationship of each segment, and obtain the steady-state interval parameters for regulation. 8.The AI-based liquid fertilizer drip irrigation regulation method for agricultural planting of claim 1, wherein, The change in flow signal refers to the fluctuation of the electrical signal output by the flow sensor as the liquid fertilizer flows in the pipeline, caused by changes in the speed and flow rate of the fertilizer solution. The irrigation node refers to the location node in the drip irrigation system that can be independently controlled and has its data collected, including the main pipe, branch pipe and zone valve.

9. An AI-based liquid fertilizer drip irrigation control system for agricultural planting, characterized in that, The system is used to implement the AI-based drip irrigation control method for liquid fertilizer in agricultural planting as described in any one of claims 1-8, and the system comprises: The feedback acquisition module is based on the liquid fertilizer storage tank, analyzes the changes in flow signals in the pipeline, detects the changes in the flow rate of the drip irrigation main pipe, monitors the response of soil moisture and ion concentration in the crop root zone, identifies the periodic signal amplitude characteristics, and obtains the feedback feature vector set. Based on the feedback feature vector set, the weight analysis module compares the performance of each parameter among different irrigation nodes, analyzes the delivery status of the drip irrigation main node, determines the node association based on the matching degree between the root zone feedback signal and the pipeline status, calculates the delivery difference, and obtains the node weight distribution information. The priority determination module judges the root zone feedback response of the drip irrigation zone based on the node weight distribution information, and analyzes the feedback trend of the drip irrigation path in conjunction with the liquid state of the mixed fertilizer pump. The node scheduling order is adjusted according to the trend to obtain the control priority sequence. Based on the control priority sequence, the action execution module optimizes the action sequence of the partitioned solenoid valves, issues valve action commands in order, records the valve action response fed back by the controller, analyzes the time difference between the command sequence and the response process, and obtains the execution timing offset signal. Based on the execution timing offset signal, the steady-state monitoring module filters continuous periodic parameter changes, analyzes the stability state of the valve response process, records the control structure state when the response is consistently consistent, and supplements the scheduling record when response fluctuations are detected, thereby obtaining the control steady-state interval parameters.

Images